A resume tells a story: five years at a recognizable company, a stack that lines up with the job description, a promotion at exactly the right moment. Screening has always rested on two assumptions about that story — that it is broadly true, and that checking it is expensive enough to postpone until the finalist stage. Generative AI broke both assumptions at the same time.
What actually changed
The cost of producing a plausible, well-targeted professional history fell to roughly zero. A candidate no longer has to invent a career badly; a model will tailor one to the posting, in the right register, with the right vocabulary, in seconds. The same tools that help an honest applicant present themselves clearly also let a dishonest one manufacture a history that survives a first read.
Application volume rose at the same moment. When every applicant can generate a bespoke application in under a minute, recruiters see more candidates per role — and per-candidate attention falls precisely when scrutiny should be rising. That is the squeeze: more submissions, better-camouflaged claims, and no additional hours in the day.
Why traditional screening cannot absorb it
Most screening stacks were designed for a world where fabrication was laborious. They break for structural reasons, not because recruiters are careless:
- Keyword matching rewards the exact thing models are good at. Fluent, on-topic, keyword-dense text is now free to produce. Ranking by it selects for generation quality, not competence.
- Background checks run too late and answer a different question. They confirm identity, criminal record, and sometimes education — usually after an offer, and rarely the substance of what the candidate claimed to have built.
- Reference checks are self-selected. A candidate nominates the people who will vouch for them. That is useful signal about relationships, and weak signal about accuracy.
- Manual vetting does not scale. Deep review of a profile is genuinely effective and genuinely expensive, so it gets rationed to the shortlist — after the compromised profile has already made it through.
The signals that still hold
Fabrication is easy at the level of prose and hard at the level of structure. A generated history is internally plausible but externally fragile, and that is where verification should look:
- Timeline coherence. Overlapping full-time engagements, tenure that does not fit the seniority claimed, gaps that appear in one artifact and not another.
- Cross-artifact consistency. The resume, the public profile, the portfolio, and the credential should describe the same person. Fabrications drift when they have to agree with each other.
- Document integrity. Certificates, transcripts, and offer letters carry structural evidence — metadata, edit traces, layout irregularities — that survives visual polish.
- Corroboration against independent sources. Claims that can only be confirmed by the claimant deserve a different weight than claims that can be checked elsewhere.
What verification-first hiring looks like
The fix is not more suspicion. It is moving verification from the end of the funnel to the front, and making it cheap enough to apply to everyone rather than rationing it to the shortlist.
Four principles make that workable:
- Verify early, at the top of the funnel. A risk signal is worth far more before four people have spent an hour interviewing.
- Score risk; do not issue verdicts. The output should be an explainable set of signals a human can inspect and overrule, not a black-box pass or fail.
- Make it fast enough to be universal. If a check takes a day, it will be used selectively — and selective checking reintroduces the bias it was meant to remove.
- Keep the human deciding, and keep the record. Verification informs the decision. An audit trail is what makes the decision defensible later.
Where WackoWave fits
WackoWave sits between the electronic document and the reality it claims to describe. Every candidate profile is analyzed by five concurrent engines — timeline, consistency, document integrity, corroboration, and anomaly detection — and consolidated into a single explainable risk score in under 30 seconds. No shortlist rationing, no week-long wait, no black box.
Matching finds candidates. Validation proves them. AI changed recruiting; trust has to catch up.